Benchmark data and code for Question-Answering on Movie stories
Switch branches/tags
Nothing to show
Clone or download
Pull request Compare This branch is 101 commits ahead, 6 commits behind makarandtapaswi:master.
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
MemN2N
MovieQA
Plotting
figures
src
test
w2v
w2v_in_progress
.gitignore
ProjectProposalMarked.pdf
README.md
correctfile.html
cross_valid_split.py
embed_plt.npz
fusiontester.py
id_map_plt.pkl
main.py
runAllTest.sh
tfIdfMatrix.obj
util.py
w2vtester.py
wrongfile.html

README.md

To run all test

bash runAllTest.sh

To run actual TFIDF file

python main.py

Requires:

Python 2.7

NLTK

pip install nltk

wordnet corpora

run nltk.download() in python interpreter and download wordnet under the corpora tab.

Plots Data

Manually get the script movieqa_get_stories.sh from movieqa.cs.toronto.edu by signing an agreement. cd MovieQA bash movieqa_get_stories.sh true

The output files are:

  • output.log
  • correctFile.html
  • wrongFile.html
  • testResults.txt

bABi dataset

sudo apt install luarocks sudo luarocks --from=https://raw.githubusercontent.com/torch/rocks/master/ install class sudo luarocks --from=https://raw.githubusercontent.com/torch/rocks/master/ install torch sudo luarocks make babitasks-scm-1.rockspec

15, 16, 17 should work, others do not work for now

babitasks 17 1

Generate 1000 training data for each of 20 tasks

for i in seq 1 20; do babi-tasks $i 1000 > task_$i.txt; done